OurGroceries MCP
Unofficial MCP server to manage grocery lists in OurGroceries. Enables listing, adding, completing, removing, and moving items.
README
<!-- mcp-name: io.github.OriginalByteMe/our-groceries-mcp -->
OurGroceries MCP
An unofficial Model Context Protocol (MCP) server for an OurGroceries account. It relies on an unofficial API client; upstream behavior, availability, and compatibility can change without notice. Review every result and verify important changes in OurGroceries.
[!WARNING] The Streamable HTTP server has no built-in application authentication. Never expose it directly to the internet: anyone who can reach it can use the configured grocery account. Bind it to localhost, or put it behind an authenticated, trusted tunnel or reverse proxy that enforces authentication before forwarding requests.
Prerequisites
- An OurGroceries account and its credentials.
- For a source install: uv and Python 3.12 or later.
- For the container option: Docker.
- An MCP client for stdio or Streamable HTTP.
The server reads OURGROCERIES_EMAIL and OURGROCERIES_PASSWORD from its process environment. .env is a local template, not an automatically loaded configuration file. Keep it private and never commit it.
Install and configure
Source with uv
git clone https://github.com/OriginalByteMe/our-groceries-mcp.git
cd our-groceries-mcp
cp .env.example .env
# Edit .env locally; do not commit it.
uv sync --locked
Set placeholders in .env only on your machine:
OURGROCERIES_EMAIL=""
OURGROCERIES_PASSWORD=""
For a shell-launched server, export the file first:
set -a
. ./.env
set +a
Docker
The image starts the Streamable HTTP transport. Build it and publish the port only on localhost:
docker build -t our-groceries-mcp .
docker run --rm --env-file .env -p 127.0.0.1:8000:8000 our-groceries-mcp
Its MCP endpoint is http://127.0.0.1:8000/mcp. This is an MCP endpoint, not a browser UI. If remote access is necessary, retain the localhost binding and use an authenticated trusted tunnel or reverse proxy; do not forward the port unauthenticated.
Install the agent skill
For agents that support Vercel Skills, install the repository skill:
npx skills add OriginalByteMe/our-groceries-mcp --skill our-groceries-mcp
This installs guidance, not credentials. Configure credentials locally in the MCP client or environment.
Connect an MCP client
Stdio (recommended)
After uv sync --locked, add this JSON to your client's MCP server configuration. Replace the absolute checkout path and placeholders locally; do not put real credentials in a shared config file.
{
"mcpServers": {
"our-groceries": {
"command": "uv",
"args": [
"run",
"--directory",
"/absolute/path/to/our-groceries-mcp",
"--locked",
"our-groceries-mcp"
],
"env": {
"OURGROCERIES_EMAIL": "",
"OURGROCERIES_PASSWORD": ""
}
}
}
}
Stdio keeps the server local to the client process and is the safest default.
Streamable HTTP
For a source checkout, after exporting the environment variables, run:
uv run --locked our-groceries-mcp \
--transport streamable-http \
--host 127.0.0.1 \
--port 8000
Connect an MCP client to http://127.0.0.1:8000/mcp. The endpoint has no HTTP authentication. Do not use 0.0.0.0 or a public host unless an authenticated trusted tunnel or reverse proxy protects every request.
Read-only smoke check
From a configured source checkout, this launches the server over stdio, initializes MCP, and calls the real read-only list_grocery_lists tool. It prints no grocery-list contents.
set -a
. ./.env
set +a
uv run --locked python - <<'PY'
import asyncio
import os
from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client
async def main() -> None:
server = StdioServerParameters(
command="uv",
args=["run", "--locked", "our-groceries-mcp"],
env=dict(os.environ),
)
async with stdio_client(server) as (read, write):
async with ClientSession(read, write) as session:
await session.initialize()
result = await session.call_tool("list_grocery_lists", {})
if result.isError:
raise RuntimeError("list_grocery_lists failed")
print("Read-only MCP smoke check passed.")
asyncio.run(main())
PY
This contacts the live OurGroceries account. Do not run it with credentials you do not intend to use.
Tools and confirmation
List and item names are resolved by case-insensitive exact match. Resolve ambiguity with list_grocery_lists or get_items before a mutation.
| Tool | Parameters | Behavior |
|---|---|---|
list_grocery_lists |
none | Read available lists. |
get_items |
list_name |
Read items in one list. |
add_item |
list_name, item, optional quantity, note |
Adds immediately; it has no confirmation flag. Ask before calling it. |
complete_item |
list_name, item, optional confirm |
Returns a preview by default. Marking complete requires confirm: true. |
remove_item |
list_name, item, optional confirm |
Returns a preview by default. Permanent removal requires confirm: true. |
move_item |
source_list, destination_list, item, optional confirm |
Returns a preview by default. Moving requires confirm: true. |
For complete_item, remove_item, and move_item, first call with the default confirm: false, show the preview, and only call again with confirm: true after explicit user approval. add_item is an immediate write, so obtain approval before the first call.
[!CAUTION]
move_itemis not atomic: it adds the item to the destination, then removes it from the source. If removal fails, the item can exist in both lists. Inspect both lists before retrying so a retry does not create another duplicate.
Troubleshooting
- Credentials are not configured: pass both environment variables to the server process;
.envalone is not loaded automatically. - Authentication failed or requests fail: verify the account credentials and network, then retry later. This is an unofficial integration and upstream API changes can break it.
- List or item not found / ambiguous: use the read tools and provide one case-insensitive exact name; rename duplicate lists or items in OurGroceries if needed.
- Cannot connect over HTTP: ensure the server is running, use
/mcp(not/), and keep the host/port local unless an authenticated proxy or tunnel is in place. - A move reports partial failure: check both source and destination lists manually before deciding whether to remove, move, or retry the item.
Development and testing
Install development dependencies and run the offline test suite with the lockfile:
uv sync --locked --all-groups
uv run --locked pytest -m "not live"
Tests marked live require a deliberately configured live account and are excluded by default. Do not add credentials to test output, fixtures, or committed files.
Security
See SECURITY.md before enabling Streamable HTTP. In particular, direct unauthenticated internet exposure grants grocery-account access to callers.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.